A new ECG arrhythmia clustering method based on Modified Artificial Bee Colony algorithm, comparison with GA and PSO classifiers

In order to diagnose the arrhythmias in electrocardiographic signals automatically, new methods for automated ECG analysis and heart beat type classification are being developed. In this paper, we proposed a new method Modified Artificial Bee Colony (MABC) algorithm for data clustering and it is applied to ECG signal analysis for arrhythmia classification. This new developed classifier based on MABC algorithm is called MABCC. The results of MABCC are compared with two other classifier's (Genetic Algorithm and Particle Swarm Optimization based) success rate results. ECG data is obtained from MITBIH database. In this study, a detailed analysis has been done on time domain features. When ECG signals are analyzed, choosing distinctive features has important effect to get a high classification success rate. By using the right features in MABC algorithm, high system classification success rate (98.73%) is achieved by MABC Classifier, similar to GA (98.59%) and PSO (99.24%). MABC has also high sensitivity for all beat types. Other methods have lower or poor classification success rates for some beat types.